Contribution
The paper gives the combination problem a structure. Before it the problem was mostly James’s subject-summing worry, Seager’s name for it and Goff’s sharpening. Chalmers splits it into subject, quality and structure problems, each with a sharp form, and adds unity, boundary, awareness and grain. He states each as an explicit argument with numbered premises. He then classifies every response by its constitutive structure, which shows that several proposed “solutions” (Rosenberg’s high-level individuals, Gabora, Seager’s infusion, IIT as a fundamental law) are emergent or identity panpsychism and do not combine anything. It does not solve the problem or prove it unsolvable, and says so (§1).
Key insight
Panpsychism was attractive because it promised a place for consciousness in physics without dualism’s problem of mental causation. That promise holds only for constitutive Russellian panpsychism, on which micro-experiences are the quiddities of physical properties and constitute macro-experience. That view is constrained to build all of our experience from a handful of micro-properties by necessary relations, with no new laws: “a little like trying to juggle seven balls in the air with both hands tied behind one’s back” (§9). Every escape from the combination problem gives up one of those constraints, and with it the motivation.
Assumptions
- Terminology from the companion paper (§2). Micro- and macrophenomenal properties; constitutive panpsychism (macrophenomenal truths are grounded in microphenomenal truths) versus non-constitutive; Russellian panpsychism (microphenomenal properties are quiddities, the categorical bases of microphysical dispositions); panprotopsychism (protophenomenal properties that are not themselves phenomenal but collectively constitute them, by a priori entailment).
- Causal closure of the microphysical. It drives every objection from mental causation (§§5.1–5.3).
- Conceivability–possibility links. Premise (2) of the subject-summing and zombie arguments. Chalmers notes that panpsychists who rejected physicalism by conceivability arguments are “not in a good position” to deny them (§4.2).
- Structure-preservation (§8, premise 4): if micro-phenomenal structure is isomorphic to micro-physical structure, anything they constitute is isomorphic to something micro-physics constitutes.
- Introspective datum (§8, premise 5): the structure of the visual field is not the structure of the brain.
Key results
All are informal arguments and classifications. There are no theorems and no data.
- The subproblems (§3).
- Subject: how do micro-subjects yield macro-subjects? Sharp form: subject-summing. Given 101 subjects, the first 100 do not seem to necessitate the 101st.
- Quality: how do micro-qualities yield macro-qualities? Sharp form: the palette problem.
- Structure: how does micro-experiential structure yield macro-experiential structure? Sharp form: structural mismatch.
- Unity, boundary (Rosenberg 1998), awareness and grain (Maxwell, Lockwood).
- A solution “must address all of these problems”, and any proposal “should indicate which problems it is addressing, and which problems it is not”.
- The panprotopsychist has an easier subject problem, since protoexperiences need no subjects. In exchange there is a new one: how non-experiences constitute experiences.
- Arguments against constitutive panpsychism (§4).
- Anti-aggregation (§4.1): from James’s nihilism about composites. Weak, because nihilism is widely rejected.
- Subject-summing (§4.2): no group of subjects necessitates a further subject, supported by a conceivability argument. Called “a significant challenge”.
- Conceivability (§4.3): a “panpsychist zombie” world, micro-physically and micro-phenomenally like ours with no macro-consciousness, is conceivable.
- Knowledge (§4.4): Mary told all micro-physical and micro-phenomenal facts. “Highly inconclusive”, since we know so little about micro-phenomenal properties.
- Palette (§4.5): against Russellian versions only.
- Revelation (§4.6): best answered by distinguishing the nature of a phenomenal property from the grounds of its instances.
- Structural mismatch (§4.7, §8): an inconsistent tetrad, later a five-premise argument. Called “one of the more powerful arguments”.
- Non-combinatorial responses (§5).
- Emergent panpsychism avoids combination and pays in mental causation. Rosenberg, Gabora and IIT read as a fundamental law are classed here (§5.1).
- Autonomous panpsychism, including cosmopsychism, faces a decomposition problem “just as hard” (§5.2).
- Identity panpsychism: the dominant-monad view, identity cosmopsychism and quantum holism. The last is the most attractive. It faces doubt about stable brain-level entanglement and a structural mismatch (§5.3).
- Combinatorial responses (§6).
- Deflating the subject: eliminativism leaves a “bundle combination problem” (§6.1).
- Combinatorial infusion (Seager) is synchronic or diachronic. If diachronic it is not constitution. On analysis it is identity or emergent panpsychism (§6.2, fn 6).
- Phenomenal bonding (Goff). A phenomenal relation could be the quiddity of spatiotemporal or causal relations. If bonding is transitive, given ubiquitous causal relations, there is “a single giant subject”. If not, “far too many subjects”. Informational integration “along the lines of Tononi’s hypothesis” might bound subjects, but it is derivative, so a new combination problem opens (§6.3).
- Deflating awareness: a functional analysis of awareness (Coleman) meets the conceivability objection to functionalism (§6.4).
- Panqualityism avoids micro-subjects but leaves all three problems open (§6.5).
- Footnote 9: Roelofs’s mereological inheritance needs strong a posteriori necessities.
- Quality (§7). Small-palette views need an account of quality combination. Large-palette views make most micro-qualities epiphenomenal or need a much more complex physics.
- Structure (§8). The best reply separates narrowly macrophysical structure (space, mass, charge) from broadly macrophysical structure (chemical, biological, computational, informational). Experience mirrors the second. Chalmers took this line in The Conscious Mind. He now says it is “perhaps the only viable line”, but it is unclear how informational composition could be necessary rather than “brute nomic principles”.
- Footnote 7, on exclusion. Nagel suggested a panprotopsychist infusion view: an isolated entity’s protophenomenal character fixes a subject, while in “the right kind of complex system” it contributes to a merged subject. Chalmers likens this to IIT’s exclusion postulate, which he glosses as “a system is conscious iff it is not part of (and does not contain) a system with a higher degree of integrated information”. “Both views seem to have the counterintuitive consequence that consciousness is extrinsic.”
- Verdict (§9). Most worth exploring: phenomenal bonding or quantum holism (subjects), small palettes (qualities), informational composition (structure), and a somewhat deflationary account of awareness. It is “not at all clear” that they can work together.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | The combination problem is several problems (subject, quality, structure, unity, boundary, awareness, grain), and a solution must address all of them | moderate | conceptual analysis §3; the “must address all” is a stipulated standard |
| C2 | Subject-summing is a significant challenge to constitutive panpsychism | moderate | conceivability argument §4.2; rests on the conceivability–possibility link |
| C3 | Structural mismatch is among the strongest arguments against constitutive Russellian pan(proto)psychism | moderate | five-premise argument §8; premise 4 is “a plausible general principle”, not shown |
| C4 | Emergent, autonomous and identity panpsychism avoid combination only at the price of mental causation or revised physics | moderate | case analysis §5, reusing the causal-exclusion argument against dualism |
| C5 | Combinatorial infusion is not combinatorial: it reduces to identity or emergent panpsychism | moderate | synchronic/diachronic dilemma §6.2, fn 6 |
| C6 | Phenomenal bonding must avoid both a universal subject and fragmentary subjects, and no fundamental relation obviously does | weak–moderate | the transitivity dilemma §6.3; intermediate options are conceded open |
| C7 | IIT’s exclusion postulate, like Nagel’s infusion view, makes consciousness extrinsic | weak (a one-line observation) | fn 7 only; no case worked through |
| C8 | Phenomenal structure most plausibly follows informational structure, but it is unclear how that can be constitutive | author’s judgement | §8; inherited from The Conscious Mind |
| C9 | The most promising avenues are phenomenal bonding or quantum holism, small palettes, informational composition and deflated awareness | author’s judgement | §9; no argument that they cohere |
Concepts
- Constitutive / Russellian panpsychism — macrophenomenal truths grounded in microphenomenal ones; microphenomenal properties as quiddities of microphysical roles (§2).
- Subject-summing problem — no group of subjects seems to necessitate a further subject (§3).
- Palette problem — few micro-qualities, many macro-qualities (§3); Dainton’s “derivation problem”.
- Structural mismatch — macrophenomenal structure is not isomorphic to macrophysical structure (§3, §8).
- Panpsychist zombie — a micro-physical and micro-phenomenal duplicate of us without macro-consciousness (§4.3).
- Emergent / autonomous / identity / combinatorial panpsychism — the four classes of response (§5).
- Quantum holism — macro-subjects identical to entangled fundamental systems (§5.3).
- Combinatorial infusion — Seager’s merging of subjects (§6.2).
- Phenomenal bonding — Goff’s phenomenal relations that constitute a new subject; co-consciousness is the natural candidate (§6.3).
- Decomposition problem — the reverse problem for cosmopsychism (§5.2).
- Narrowly / broadly macrophysical structure — physics-level versus any structure micro-physics constitutes, information included (§8).
Connections
James’s Principles of Psychology is the origin. Chalmers’s epigraph is its twelve-men passage, which ends that “the private minds do not agglomerate into a higher compound mind”. That is the group-mind argument Schwitzgebel’s paper (LIT-159, NOTE-131) runs against. Schwitzgebel cites this paper only in his bibliography and fn 25 (NOTE-131, Composition). Goff 2009 supplies the argument forms of §4, and Seager, Coleman, Rosenberg, Roelofs and Mørch the responses. The companion paper, Panpsychism and Panprotopsychism, supplies the terminology and the case for panpsychism; this one is the case against. That companion is among the core consciousness works being filed alongside this reading. The conceivability argument against physicalism, from Chalmers’s core hard-problem work, is the template for §4.3. IIT appears twice: as emergent panpsychism (§5.1) and, via the exclusion postulate, in fn 7. The record holds IIT 3.0 (LIT-445, NOTE-365) and a sympathetic assessment of its panpsychism (LIT-185, NOTE-122).
Bearing on the record
- THEORY-043: supports its structure, adds one argument.
- James’s argument as the ancestor of anti-nesting. Subject-summing’s premise is that no group of subjects necessitates a further subject. It would refuse the upward move to a minded nation as firmly as Putnam’s stipulation. But it is aimed at constitutive panpsychism, where the whole’s experience is grounded in the parts’ experiences. The account’s functionalist premise grounds it in organisation instead, so the argument does not refuse Schwitzgebel’s move as stated.
- Footnote 7 (C7) is an objection to exclusion that owes nothing to the verdict on groups: exclusion makes consciousness extrinsic. The account says exclusion is a postulate justified weakly by the verdict it gives. This adds that it carries a cost of its own. It is one line, not an argument worked through.
- The boundary problem is the account’s seam in panpsychist dress. Bonding must bound a human subject without bonding it to everything causally connected. The candidate Chalmers names, Tononi-style informational integration, is the account’s architectural route, and he finds it derivative and unprincipled. So the paper arrives, independently, at the account’s two routes and at its verdict that neither is yet principled. It does not discuss groups.
- LIT-439 / NOTE-341 (Levin), C8. Levin’s “Chalmers 2013” is Panpsychism and Panprotopsychism (his reference list, PMC6923654), not this paper. Measured against this paper’s standard (C1), his remark that the model “takes a stand on the combination problem” names no sub-problem and no constitution relation, so it is not yet a stand. His nested Selves bear on the boundary problem for agency. For experience they would need a functional account of awareness, which §6.4 argues faces the conceivability objection. NOTE-341’s “What the stand is is not said” is accurate.
- NOTE-131’s correspondence principle is not here. Nothing in this paper says a conscious system’s complex capacities must arise largely from relations among subsystems rather than their capacities. The nearest idea is §8’s informational composition, which concerns phenomenal structure, not where capacities come from.
- No ML instruction; nothing for the anthology.
Limitations
- A map, not a result. No solution and no impossibility proof (§1).
- The verdicts are judgements of promise. §9’s shortlist is not shown to cohere.
- Scope. It treats panpsychism and panprotopsychism; materialist and dualist composition are out of scope. Groups of people appear only in the James epigraph.
- Draft, not print. Read from the April 2015 draft. Footnotes say material was added after the first draft, so the published chapter may add or move passages, including fn 7 and fn 9.
- Conceivability premises. Several arguments stand or fall with the conceivability–possibility link. The paper defends it only ad hominem against panpsychists.
Open questions
- Is there a bonding relation that is “just nontransitive enough” to yield bounded macro-subjects (§6.3)? Any answer would also be a criterion for which wholes are subjects, and so would bear on whether a nation is one.
- Can informational composition be made constitutive rather than a brute law (§8)?
- Does the extrinsicness objection (fn 7) survive IIT 4.0’s restatement of exclusion? It is aimed at the 2008-era postulate.
Corrections
- No seeded skim to correct.
- To the brief. Levin’s citation of “Chalmers 2013” is to Panpsychism and Panprotopsychism, not this paper. The 2013 circulation date for this paper could not be confirmed: the posted draft is April 2015, and the paper began as a July 2012 talk. The book’s year is 2017; Crossref dates it 29 December 2016.
- Slips in the draft. §5 says “The first three are noncombinatorial responses, while the third is a combinatorial response”; the fourth is meant. The reference list cites James’s Principles as 1895; it was published in 1890.